Intelligent positioning method insensitive to image changes

Abstract

The application discloses an intelligent positioning method which is not sensitive to image changes, which comprises the following steps: flattening a pre-recorded environment image and constructing a coordinate system on the flattened environment image; building an identification and positioning unit and performing big data learning; constructing a basic skeleton and a basic outer frame of an identification target model in the identification and positioning unit; when a target enters a positioning monitoring area of a visual monitoring device, the visual monitoring device captures a skeleton configuration of the moving target, compares the obtained skeleton configuration with the basic skeleton to determine whether the target conforms to the type of the identification target model, and when the target conforms to the type of the identification target model, marks a target outer frame on the target, and meanwhile, ignores image changes in the part outside the target frame in the background image; when the target outer frame moves to the edge of the positioning monitoring area and the target outer frame is shielded, the visual monitoring device stops tracking the target; and the application has the advantages that the target can be quickly locked and low-weight image changes can be ignored.

Description

Technical Field

[0001] The invention relates to the field of monitoring and positioning, and in particular to an intelligent positioning method which is insensitive to image changes. Background Art

[0002] Image recognition refers to the use of computers to process, analyze, and interpret images to identify various patterns of objects. In general industrial applications, industrial cameras are used to capture images, and then software is used to perform further recognition and processing based on grayscale differences. Representative image recognition software companies include Cognex and others abroad, and Image Intelligence in China. In geography, it also refers to the technology used to classify remote sensing images.

[0003] Based on image recognition technology, an on-site positioning monitoring system with direct on-site image recognition has been established. For on-site monitoring, the monitoring agency needs to be able to lock on to the monitored target and eliminate interference caused by non-critical image changes in the on-site environment. To achieve this goal, we need an intelligent positioning method that is insensitive to image changes. Summary of the Invention

[0004] Purpose of the invention: The purpose of the present invention is to address the deficiencies of the prior art and provide an intelligent positioning method that is insensitive to image changes.

[0005] Technical solution: The present invention provides an intelligent positioning method that is insensitive to image changes, comprising the following steps:

[0006] Step 1: Define a specified positioning monitoring area, set up a matching visual monitoring device for the corresponding positioning monitoring area, use the visual monitoring device to pre-record the environmental status of the positioning monitoring area in an open state, flatten the pre-recorded environmental image, and construct a coordinate system on the flattened environmental image;

[0007] Step 2: Build a recognition and positioning unit. Pre-enter the recognition target model and use a crawler network to crawl images that match the recognition target model type on the public network. The recognition and positioning unit then performs big data learning based on the crawled images. Through big data learning, the basic skeleton and basic outer framework of the recognition target model are constructed in the recognition and positioning unit.

[0008] Step 3: When a target enters the positioning monitoring area of ​​the visual monitoring device, the visual monitoring device captures the skeleton structure of the active target and compares the obtained skeleton structure with the basic skeleton to determine whether the target meets the type of the recognition target model;

[0009] Step 4: When the target meets the type of the identification target model, the target outer frame is marked on the target by the identification positioning unit according to the basic outer frame, and the visual monitoring device is controlled by the identification positioning unit to point tracking the target outer frame, while ignoring the image changes in the background image outside the target outer frame in the target monitoring block.

[0010] Step 5: When the target outer frame moves to the edge of the positioning monitoring area and the target outer frame is shielded, the visual monitoring device stops tracking the target.

[0011] As preferred, in step 1, the planarized environment image is divided into several monitoring blocks while establishing the coordinate system.

[0012] As preferred, in step 4, the visual monitoring device is controlled by the identification positioning unit to point tracking the target outer frame, while ignoring the image changes in the background image outside the target frame in the target monitoring block.

[0013] As preferred, in step 4, the target outer frame is marked on the target, and the coordinate position of the target outer frame in the coordinate system is marked and the change of the coordinate position of the target outer frame is tracked in real time.

[0014] As preferred, in step 5, when the target outer frame moves to the edge of the positioning monitoring area and at least two-thirds of the target outer frame is shielded, the visual monitoring device stops tracking the target.

[0015] As preferred, in step 3, when the obtained skeleton structure is compared with the basic skeleton, a proportion parameter is introduced, and when the proportion parameter is more than four times up and down compared with the basic skeleton, even if the skeleton structure is the same as the basic skeleton, the target to be identified is not considered to meet the type of the identification target model.

[0016] The present application has the following beneficial effects compared with the prior art: The pre-recorded environment image is planarized and the coordinate system is constructed on the planarized environment image, the basic skeleton and the basic outer frame of the identification target model are constructed in the identification positioning unit through big data learning, the target to be identified and positioned entering the positioning monitoring area can be quickly identified by comparing the skeleton structure of the target with the basic skeleton, and the target outer frame is marked on the target, so as to point tracking the target, while ignoring the image changes in the background image outside the target outer frame, so that the unimportant changes in the background image environment will not affect the positioning accuracy of the target. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present application will be clearly and completely described below. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of the present application.

[0018] In the description of the present application, it should be understood that the orientation or positional relationship indicated by the terms "center", "longitudinal", "transverse", "upper", "lower", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "axial", "radial", "circumferential" and the like is for the convenience of describing the present application and simplifying the description, and does not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation of the present application.

[0019] In the present application, unless otherwise explicitly specified and limited, the terms "mounting", "connection", "connecting", "fixing" and the like should be understood in a broad sense, for example, can be fixed connection, can also be detachable connection, can also be integrated; can be mechanical connection, can also be electrical connection, can also be communication connection; can be direct connection, can also be indirect connection through intermediate medium, can be the internal communication of two elements or the interaction relationship of two elements, unless otherwise explicitly limited. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0020] The technical solutions of the present application will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes can not be described in some embodiments.

[0021] Embodiment 1: An intelligent positioning method insensitive to image changes, comprising the following steps:

[0022] Step 1: Define a specified positioning monitoring area, erect a matching visual monitoring device for the corresponding positioning monitoring area, use the visual monitoring device to pre-record the environment state of the positioning monitoring area in the open state, and flatten the pre-recorded environment image and construct a coordinate system on the flattened environment image;

[0023] Step 2: Build an identification positioning unit, pre-input an identification target model and use a crawler network to capture images conforming to the identification target model type on a public network, and based on the captured images, the identification positioning unit performs big data learning, and through big data learning, the identification positioning unit constructs a basic skeleton and a basic outer frame of the identification target model;

[0024] Step 3: When a target enters the positioning monitoring area of the visual monitoring device, the visual monitoring device captures the skeleton configuration of the moving target, and compares the obtained skeleton configuration with the basic skeleton to determine whether the target conforms to the type of the identified target model;

[0025] Step 4: When the target conforms to the type of the identified target model, the visual monitoring device marks the target outer frame on the target according to the basic outer frame through the identification positioning unit, and the visual monitoring device is controlled by the identification positioning unit to point tracking the target outer frame, while ignoring the image changes in the part outside the target frame in the target monitoring block in the background image;

[0026] Step 5: When the target outer frame moves to the edge of the positioning monitoring area and the target outer frame is shielded, the visual monitoring device stops tracking the target.

[0027] Embodiment 2: An intelligent positioning method insensitive to image changes, comprising the following steps:

[0028] Step 1: Define a specified positioning monitoring area, erect a matching visual monitoring device for the corresponding positioning monitoring area, use the visual monitoring device to pre-record the environment state of the positioning monitoring area in the empty state, and flatten the pre-recorded environment image and construct a coordinate system on the flattened environment image. At the same time of establishing the coordinate system, the flattened environment image is divided into several monitoring blocks;

[0029] Step 2: Build an identification positioning unit, pre-input an identification target model, and use a crawler network to capture images conforming to the type of the identification target model on a public network, and based on the captured images, the identification positioning unit is learned by big data, and the basic skeleton and the basic outer frame of the identification target model are constructed in the identification positioning unit through big data learning;

[0030] Step 3: When a target enters the positioning monitoring area of the visual monitoring device, the visual monitoring device captures the skeleton configuration of the moving target, and compares the obtained skeleton configuration with the basic skeleton to determine whether the target conforms to the type of the identified target model;

[0031] Step 4: When the target conforms to the type of the identified target model, the visual monitoring device marks the target outer frame on the target according to the basic outer frame through the identification positioning unit, and the visual monitoring device is controlled by the identification positioning unit to point tracking the target outer frame, while ignoring the image changes in the part outside the target frame in the target monitoring block in the background image;

[0032] Step 5: When the target outer frame moves to the edge of the positioning monitoring area and at least two-thirds of the target outer frame is shielded, the visual monitoring device stops tracking the target.

[0033] Example 3: An intelligent positioning method that is insensitive to image changes, comprising the following steps:

[0034] Step 1: Define a specified positioning monitoring area, set up a matching visual monitoring device for the corresponding positioning monitoring area, use the visual monitoring device to pre-record the environmental status of the positioning monitoring area in an open state, flatten the pre-recorded environmental image, and construct a coordinate system on the flattened environmental image;

[0035] Step 2: Build a recognition and positioning unit. Pre-enter the recognition target model and use a crawler network to crawl images that match the recognition target model type on the public network. The recognition and positioning unit then performs big data learning based on the crawled images. Through big data learning, the basic skeleton and basic outer framework of the recognition target model are constructed in the recognition and positioning unit.

[0036] Step 3: When a target enters the positioning monitoring area of ​​the visual monitoring device, the visual monitoring device captures the skeletal structure of the moving target and compares the obtained skeletal structure with the basic skeleton to determine whether the target meets the type of recognition target model. When comparing the obtained skeletal structure with the basic skeleton, a ratio parameter is introduced. If the ratio parameter of the skeletal structure is more than four times that of the basic skeleton, the target to be identified is not considered to meet the type of recognition target model, even if the skeletal structure and the basic skeleton have the same configuration;

[0037] Step 4: When the target meets the type of the recognition target model, the recognition and positioning unit marks the target outer frame according to the basic outer frame. The visual monitoring device tracks the target outer frame at a fixed point under the control of the recognition and positioning unit. At the same time, the coordinate position of the target outer frame is marked in the coordinate system and the changes of the coordinate position of the target outer frame are tracked in real time, while ignoring the image changes of the part of the background image outside the target outer frame.

[0038] Step 5: When the outer frame of the target moves to the edge of the positioning monitoring area and at least two-thirds of the outer frame of the target is blocked, the visual monitoring device stops tracking the target.

[0039] The advantage of this technical solution is that it flattens the pre-recorded environmental image and constructs a coordinate system on the flattened environmental image. Through big data learning, the basic skeleton and basic external frame of the recognition target model are constructed in the recognition and positioning unit. By comparing the target's skeleton structure with the basic skeleton, the target to be identified and positioned entering the positioning monitoring area can be quickly identified, and the target external frame can be marked for the target to track the target at a fixed point. At the same time, the image changes in the background image outside the target external frame are ignored, so that unimportant changes in the background image environment will not affect the target positioning accuracy.

[0040] In the present application, unless otherwise explicitly specified and limited, the first feature is "on" or "under" the second feature can be that the first feature and the second feature are in direct contact, or the first feature and the second feature are in indirect contact through an intermediate medium. Moreover, the first feature is "above", "over" and "on" the second feature can be that the first feature is directly above or obliquely above the second feature, or only means that the first feature is higher than the second feature in horizontal height. The first feature is "under", "below" and "under" the second feature can be that the first feature is directly below or obliquely below the second feature, or only means that the first feature is lower than the second feature in horizontal height. In the description of the present application, the description of the reference terms "one embodiment", "some embodiments", "example", "specific example" or "some examples" and the like means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are contained in at least one embodiment or example of the present application. In the present application, the illustrative description of the above-mentioned terms does not necessarily refer to the same embodiment or example.

[0041] Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, the skilled person in the art can combine and combine the different embodiments or examples described in the present application and the features of the different embodiments or examples without contradiction.

[0042] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. An intelligent positioning method that is insensitive to image changes, characterized by: The following steps are involved: Step 1: Define a specified positioning monitoring area, set up a matching visual monitoring device for the corresponding positioning monitoring area, use the visual monitoring device to pre-record the environmental status of the positioning monitoring area in an open state, flatten the pre-recorded environmental image, and construct a coordinate system on the flattened environmental image; Step 2: Build a recognition and positioning unit. Pre-enter the recognition target model and use a crawler network to crawl images that match the recognition target model type on the public network. The recognition and positioning unit then performs big data learning based on the crawled images. Through big data learning, the basic skeleton and basic outer framework of the recognition target model are constructed in the recognition and positioning unit. Step 3: When a target enters the positioning monitoring area of ​​the visual monitoring device, the visual monitoring device captures the skeleton structure of the active target and compares the obtained skeleton structure with the basic skeleton to determine whether the target meets the type of the recognition target model; Step 4: When the target meets the type of the recognition target model, the recognition and positioning unit marks the target outer frame according to the basic outer frame. The visual monitoring device tracks the target outer frame at a fixed point under the control of the recognition and positioning unit, while ignoring the image changes of the background image outside the target outer frame. Step 5: When the outer frame of the target moves to the edge of the positioning monitoring area and is blocked, the visual monitoring device stops tracking the target.

2. The intelligent positioning method insensitive to image changes according to claim 1, characterized in that: In step 1, a coordinate system is established and the planar environment image is divided into several monitoring blocks.

3. The intelligent positioning method insensitive to image changes according to claim 2, characterized in that: In step 4, the visual monitoring device tracks the target's outer frame point under the control of the recognition and positioning unit, while ignoring the image changes of the part outside the target frame in the monitoring block where the target is located in the background image.

4. The intelligent positioning method insensitive to image changes according to claim 1, characterized in that: In step 4, the target outer frame is marked for the target, and the coordinate position of the target outer frame is marked in the coordinate system and the change of the coordinate position of the target outer frame is tracked in real time.

5. The intelligent positioning method insensitive to image changes according to claim 1, characterized in that: In step 5, when the outer frame of the target moves to the edge of the positioning monitoring area and at least two-thirds of the outer frame of the target is blocked, the visual monitoring device stops tracking the target.

6. The intelligent positioning method insensitive to image changes according to claim 1, characterized in that: In step 3, when the obtained skeleton structure is compared with the basic skeleton, a scale parameter is introduced. When the scale parameter of the skeleton structure is more than four times higher than that of the basic skeleton, the target to be identified is not considered to meet the type of the identification target model even if the skeleton structure and the basic skeleton have the same configuration.